Pre-match prediction for pattern testing

Abstract

A system and method for improving speed of automated pattern testing for any particular hardware configuration. A pattern testing includes a matching process and a pre-match process that establishes features of interest used in the match process. The pre-match process is used to predict a likelihood of a passed match (or a failed match) and alters the subsequent match process in response to the prediction.

What is claimed as new and desired to be protected by Letters Patent of the United States is: 
     
       1. A machine-implemented pattern testing method comparing a first digital representation of a first pattern against a second digital representation of a second pattern using a match process for determining a match result from a first set of match features of interest of the first digital representation and a second set of match features of interest of the second digital representation using a set of predetermined match criteria, comprising executing on a processor the steps of:
 a) extracting, using the processor, a first set of features of interest from the first digital representation; 
 b) extracting, using the processor, a second set of features of interest from the second digital representation; 
 c) establishing, using the processor, the first set of match features of interest of the first digital representation from said first set of features of interest; 
 d) establishing, using the processor, the second set of match features of interest of the second digital representation from said second set of features of interest; and 
 e) predicting, using the processor, a likelihood of a passed match for the match result, responsive to an evaluation of the sets of match features of interest prior to an execution of the match process, using the sets of match features of interest; 
 
       wherein predicting the likelihood of a passed match comprises, predicting a degree of pass-match through non-match with confidence to determine whether to execute a normal match process. 
     
     
       2. A machine-implemented pattern testing method comparing a first digital representation of a first pattern against a second digital representation of a second pattern using a match process for determining a match result from a first set of match features of interest of the first digital representation and a second set of match features of interest of the second digital representation using a set of predetermined match criteria, comprising executing on a processor the steps of:
 a) extracting a first set of features of interest from the first digital representation; 
 b) extracting a second set of features of interest from the second digital representation; 
 c) establishing the first set of match features of interest of the first digital representation from said first set of features of interest; 
 d) establishing the second set of match features of interest of the second digital representation from said second set of features of interest; and 
 e) predicting a likelihood of a passed match for the match result, responsive to an evaluation of the sets of match features of interest prior to an execution of the match process, using the sets of match features of interest; further comprising: 
 f) adjusting, responsive to said likelihood of said passed match for the match result, a match criterion of the set of predetermined match criteria of the match process to produce an adjusted set of match criteria different from the set of predetermined match criteria; and 
 g) performing the match process to produce the match result using said adjusted set of match criteria instead of the set of predetermined match criteria; 
 wherein predicting the likelihood of a passed match comprises, predicting a degree of pass-match through non-match with confidence to determine whether to execute a normal match process. 
 
     
     
       3. The pattern testing method of  claim 1  further comprising:
 f) skipping the match process when said likelihood of said passed match for the match result exceeds a predetermined high likelihood threshold; and 
 g) setting the match result as said passed match without performing the match process. 
 
     
     
       4. The pattern testing method of  claim 1  further comprising:
 f) skipping the match process when said likelihood of said passed match for the match result falls below a predetermined low likelihood threshold; and 
 g) setting the match result as a failed match without performing the match process. 
 
     
     
       5. The pattern testing method of  claim 2  further comprising:
 h) skipping the match process when said likelihood of said passed match for the match result exceeds a predetermined high likelihood threshold; and 
 i) setting the match result as said passed match without performing the match process. 
 
     
     
       6. The pattern testing method of  claim 2  further comprising:
 h) skipping the match process when said likelihood of said passed match for the match result falls below a predetermined low likelihood threshold; and 
 i) setting the match result as a failed match without performing the match process. 
 
     
     
       7. The pattern testing method of  claim 1  wherein said establishing step c) and said establishing step d) includes:
 f) comparing the sets of features of interest to establish a pair of sets of conforming features of interest having a conformance metric responsive to a predetermined threshold, said pair of sets of conforming features of interest including a first set of conforming features of interest from said first set of features of interest and including a second set of conforming features of interest from said second set of features of interest wherein said sets of conforming features of interest each include a plurality of conforming features of interest common to both said first set of features of interest and to said second set of features of interest; 
 g) evaluating a group metric for said sets of conforming features of interest; 
 h) selecting said first set of conforming feature of interest as the first set of match features of interest of the first digital representation when said group metric exceeds a predetermined threshold; and 
 i) selecting said second set of conforming feature of interest as the second set of match features of interest of the second digital representation when said group metric exceeds said predetermined threshold. 
 
     
     
       8. The pattern testing method of  claim 7  wherein said predicting step e) includes an evaluation of a degree of a conformance of said sets of features of interest and said group metric when establishing said likelihood of the match result. 
     
     
       9. A non-transitory computer-readable medium for pattern testing comparing a first digital representation of a first pattern against a second digital representation of a second pattern using a match process for determining a match result from a first set of match features of interest of the first digital representation and a second set of match features of interest of the second digital representation using a set of predetermined match criteria, comprising instructions stored thereon, that when executed on a processor, perform the steps of:
 a) extracting a first set of features of interest from the first digital representation; 
 b) extracting a second set of features of interest from the second digital representation; 
 c) establishing the first set of match features of interest of the first digital representation from said first set of features of interest; 
 d) establishing the second set of match features of interest of the second digital representation from said second set of features of interest; and 
 e) predicting a likelihood of a passed match for the match result, responsive to an evaluation of the sets of match features of interest prior to an execution of the match process, using the sets of match features of interest; 
 wherein predicting the likelihood of a passed match comprises, predicting a degree of pass-match through non-match with confidence to determine whether to execute a normal match process. 
 
     
     
       10. A non-transitory computer-readable medium for pattern testing comparing a first digital representation of a first pattern against a second digital representation of a second pattern using a match process for determining a match result from a first set of match features of interest of the first digital representation and a second set of match features of interest of the second digital representation using a set of predetermined match criteria, comprising instructions stored thereon, that when executed on a processor, perform the steps of:
 a) extracting a first set of features of interest from the first digital representation; 
 b) extracting a second set of features of interest from the second digital representation; 
 c) establishing the first set of match features of interest of the first digital representation from said first set of features of interest; 
 d) establishing the second set of match features of interest of the second digital representation from said second set of features of interest; and 
 e) predicting a likelihood of a passed match for the match result, responsive to an evaluation of the sets of match features of interest prior to an execution of the match process, using the sets of match features of interest; 
 
       wherein the instructions, when executed on the processor, further comprise the steps of:
 f) adjusting, responsive to said likelihood of said passed match for the match result, a match criterion of the set of predetermined match criteria of the match process to produce an adjusted set of match criteria different from the set of predetermined match criteria; and 
 g) performing the match process to produce the match result using said adjusted set of match criteria instead of the set of predetermined match criteria; 
 wherein predicting the likelihood of a passed match comprises, predicting a degree of pass-match through non-match with confidence to determine whether to execute a normal match process. 
 
     
     
       11. The medium of  claim 9  wherein the instructions, when executed on the processor, further comprise the steps of:
 f) skipping the match process when said likelihood of said passed match for the match result exceeds a predetermined high likelihood threshold; and 
 g) setting the match result as said passed match without performing the match process. 
 
     
     
       12. The medium of  claim 9  wherein the instructions, when executed on the processor, further comprise the steps of:
 f) skipping the match process when said likelihood of said passed match for the match result falls below a predetermined low likelihood threshold; and 
 g) setting the match result as a failed match without performing the match process. 
 
     
     
       13. The medium of  claim 10  wherein the instructions, when executed on the processor, further comprise the steps of:
 h) skipping the match process when said likelihood of said passed match for the match result exceeds a predetermined high likelihood threshold; and 
 i) setting the match result as said passed match without performing the match process. 
 
     
     
       14. The medium of  claim 10  wherein the instructions, when executed on the processor, further comprise the steps of:
 h) skipping the match process when said likelihood of said passed match for the match result falls below a predetermined low likelihood threshold; and 
 i) setting the match result as a failed match without performing the match process. 
 
     
     
       15. A system for pattern testing to compare a first digital representation of a first pattern against a second digital representation of a second pattern using a match process for determining a match result from a first set of match features of interest of the first digital representation and a second set of match features of interest of the second digital representation using a set of predetermined match criteria, comprising: a processor that
 a) extracts a first set of features of interest from the first digital representation, 
 b) extracts a second set of features of interest from the second digital representation, 
 c) establishes the first set of match features of interest of the first digital representation from said first set of features of interest, and 
 d) establishes the second set of match features of interest of the second digital representation from said second set of features of interest; and a memory for storing the digital representations, sets of features of interest, and sets of match features of interest; wherein said processor further 
 e) predicts a likelihood of a passed match for the match result, responsive to an evaluation of the sets of match features of interest prior to an execution of the match process, using the sets of match features of interest; 
 wherein predicting the likelihood of a passed match comprises, predicting a degree of pass-match through non-match with confidence to determine whether to execute a normal match process. 
 
     
     
       16. A system for pattern testing to compare a first digital representation of a first pattern against a second digital representation of a second pattern using a match process for determining a match result from a first set of match features of interest of the first digital representation and a second set of match features of interest of the second digital representation using a set of predetermined match criteria, comprising: a processor that
 a) extracts a first set of features of interest from the first digital representation, 
 b) extracts a second set of features of interest from the second digital representation, 
 c) establishes the first set of match features of interest of the first digital representation from said first set of features of interest, and 
 d) establishes the second set of match features of interest of the second digital representation from said second set of features of interest; and a memory for storing the digital representations, sets of features of interest, and sets of match features of interest; wherein said processor further 
 e) predicts a likelihood of a passed match for the match result, responsive to an evaluation of the sets of match features of interest prior to an execution of the match process, using the sets of match features of interest; 
 wherein predicting the likelihood of a passed match comprises, predicting a degree of pass-match through non-match with confidence to determine whether to execute a normal match process. 
 
     
     
       17. The system of  claim 15  wherein said processor further:
 f) skips the match process when said likelihood of said passed match for the match result exceeds a predetermined high likelihood threshold; and 
 g) sets the match result as said passed match without performing the match process. 
 
     
     
       18. The system of  claim 15  wherein said processor further:
 f) skips the match process when said likelihood of said passed match for the match result falls below a predetermined low likelihood threshold; and 
 g) sets the match result as a failed match without performing the match process. 
 
     
     
       19. The system of  claim 16  wherein said processor further:
 h) skips the match process when said likelihood of said passed match for the match result exceeds a predetermined high likelihood threshold; and 
 i) sets the match result as said passed match without performing the match process. 
 
     
     
       20. The system of  claim 16  wherein said processor further:
 h) skips the match process when said likelihood of said passed match for the match result falls below a predetermined low likelihood threshold; and 
 i) sets the match result as a failed match without performing the match process.